Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations Westi on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

How to read 'top' output

Status
Not open for further replies.

kukuluku

MIS
May 2, 2002
56
US
Can someone shed some light here? How do you read the output from 'top' session? Especially memory part, real, virtual, free and page. How is my system doing in terms of resourse usage?

Thanks much,


System: hpdev Fri May 17 11:20:12 2002
Load averages: 0.90, 1.04, 1.10
254 processes: 244 sleeping, 10 running
Cpu states:
CPU LOAD USER NICE SYS IDLE BLOCK SWAIT INTR SSYS
0 0.86 4.0% 1.2% 0.8% 94.1% 0.0% 0.0% 0.0% 0.0%
1 0.94 4.2% 0.4% 3.2% 92.3% 0.0% 0.0% 0.0% 0.0%
--- ---- ----- ----- ----- ----- ----- ----- ----- -----
avg 0.90 4.2% 0.8% 2.0% 93.1% 0.0% 0.0% 0.0% 0.0%

Memory: 477092K (274904K) real, 322416K (162404K) virtual, 637232K free Page# 1/14

CPU TTY PID USERNAME PRI NI SIZE RES STATE TIME %WCPU %CPU COMMAND
0 ? 13314 oracle 154 20 28104K 1544K sleep 0:13 3.49 3.48 oraclefscnv
1 ? 12282 oracle 154 20 28232K 1688K sleep 5:25 1.88 1.87 oraclehrdemo8
0 ? 12272 oracle 154 20 28104K 1608K sleep 5:28 1.72 1.72 oraclehrdemo8
0 ? 12262 oracle 154 20 28104K 1592K sleep 5:27 1.70 1.70 oraclehrdemo8
1 ? 4967 oracle 154 20 30040K 3496K sleep 8:35 1.67 1.66 oraclehrdev
 
Not sure how on UX

but common forms are

Total (Active)
or
Total(Free)
 
That's not really enough to assess how stressed your system is. Could you post about ten lines of the output of:

'vmstat 1'

please?
Mike
________________________________________________________________________________

"Experience is the comb that Nature gives us, after we are bald."

Is that a haiku?
I never could get the hang
of writing those things.
 
Here's 'vmstat 1':
hpdev_oracle_fsdemo8 => vmstat 1
procs memory page faults cpu
r b w avm free re at pi po fr de sr in sy cs us sy id
2 0 0 41652 155647 11 6 0 0 7 0 0 754 1820 238 11 4 85
2 0 0 41652 155596 15 3 1 0 0 0 0 689 2911 319 4 0 96
2 0 0 41652 155596 12 2 1 0 0 0 0 641 2384 269 0 0 100
2 0 0 41652 155596 10 1 1 0 0 0 0 605 2018 231 12 5 83
2 0 0 41652 155935 26 2 1 0 12 0 0 570 2318 222 8 2 90
2 0 0 42471 155935 21 1 1 0 9 0 0 551 1930 192 0 1 99
2 0 0 42471 155935 17 0 1 0 7 0 0 544 1723 182 9 4 87
2 0 0 42471 155935 17 3 1 0 8 0 0 626 2482 265 5 2 93
2 0 0 42471 155935 14 2 0 0 6 0 0 596 2082 232 0 0 100
2 0 0 42471 155936 17 1 0 0 4 0 0 564 1925 205 3 2 95
4 0 0 42958 155935 14 0 0 0 3 0 0 590 1937 219 7 7 86
4 0 0 42958 155935 11 0 0 0 2 0 0 672 2529 293 10 4 86
4 0 0 42958 155935 8 0 0 0 1 0 0 709 2761 321 3 2 95



This is from top
System: hpdev Mon May 20 15:06:04 2002
Load averages: 1.12, 1.11, 1.08
253 processes: 243 sleeping, 10 running
Cpu states:
CPU LOAD USER NICE SYS IDLE BLOCK SWAIT INTR SSYS
0 1.13 7.9% 0.0% 1.6% 90.5% 0.0% 0.0% 0.0% 0.0%
1 1.10 6.3% 0.2% 1.8% 91.7% 0.0% 0.0% 0.0% 0.0%
--- ---- ----- ----- ----- ----- ----- ----- ----- -----
avg 1.12 7.1% 0.2% 1.6% 91.1% 0.0% 0.0% 0.0% 0.0%

Memory: 471876K (298740K) real, 317404K (187104K) virtual, 621572K free Page# 1/14

CPU TTY PID USERNAME PRI NI SIZE RES STATE TIME %WCPU %CPU COMMAND
1 ? 27461 oracle 154 20 28104K 1592K sleep 9:18 1.87 1.86 oraclehrdemo8
0 ? 27519 oracle 154 20 28232K 1688K sleep 9:05 1.37 1.37 oraclehrdemo8
0 ? 27498 oracle 154 20 28104K 1608K sleep 9:11 1.32 1.32 oraclehrdemo8
0 ? 32 root 152 20 0K 0K run 11:23 1.20 1.20 vxfsd
0 ? 1636 oracle 154 20 30008K 3448K sleep 5:42 1.17 1.17 oraclehrdev
0 ? 20326 oracle 154 20 30040K 3496K sleep 12:04 1.13 1.13 oraclehrdev
1 ? 19497 oracle 156 20 28456K 1972K sleep 6:15 1.05 1.05 ora_lgwr_hrdemo8
0 ? 12669 oracle 154 20 32344K 5800K sleep 1:07 0.98 0.98 oraclefscnv
1 ? 1761 root -16 20 5768K 5224K run 75:19 0.80 0.79 midaemon
1 pty/ttyp4 15382 oracle 189 24 1060K 596K run 0:00 0.88 0.65 top
1 ? 15381 oracle 154 20 28248K 1752K sleep 0:00 0.74 0.59 oraclefscnv
1 ? 27672 oracle 154 20 33640K 7064K sleep 8:26 0.51 0.51 oraclefscnv
0 ? 496 root 154 20 32K 132K sleep 52:51 0.50 0.50 syncer
1 ? 14952 oracle 154 20 30008K 3492K sleep 0:00 0.45 0.45 oraclefscnv
1 ? 12814 oracle 154 20 31848K 5336K sleep 0:17 0.38 0.38 oraclefscnv
0 ? 20010 psft 168 20 4444K 13348K sleep 1:41 0.36 0.36 PSPRCSRV


Thanks much,
 
If you look at the 'po' (page-out) column above, you'll see that it's always zero. Also -- the 'free' column doesn't change much. Yous system is not short of physical memory.

There are no processes (b)locked and no (w)aiting, the CPU idle figure figure is high but fluctuates.

So -- your system is not i/o or memory bound and shows no sign of becoming so if the load were to increase but not vary in type (more of the same kind of load)

CPU is loaded, some anyway - your system is far from being CPU bound.

Are you seeing performance problems? Mike
________________________________________________________________________________

"Experience is the comb that Nature gives us, after we are bald."

Is that a haiku?
I never could get the hang
of writing those things.
 
When user called and says their job runs extremely slow, the system doesn't seem very busy. The 'sar' shows high %idle too. I am wondering if the system is utilized it's resources smartly enough. Wonder if I should bump up memory allocation for database.

Thanks so much,
 
Also, is there a way to tell how much RAM you have for the machine from the 'top' utility? I can't seem to see it from the output.

 
I would steer clear of Top really, I don't find it useful except for finding runaway processes, the two utilities I use most are:

vmstat -- as above, good for cpu and virtual memory usage

iostat -- as below, good for disk activity

'iostat 1'

displays disk activity averaged over the last second Mike
________________________________________________________________________________

"Experience is the comb that Nature gives us, after we are bald."

Is that a haiku?
I never could get the hang
of writing those things.
 
kukuluku,

Have your user tell you when they are running the job that seems slow. While they are running it, running vmstat and iostat for several minutes.

If your system still shows that it isn't overused, it may be another problem: network problems between the user's system and your unix system, printer problems if the user considers the end of the job a hard copy printout, etc.

I worked in a place once where users complained about jobs not printing fast enough. It always happened on the last Friday of the month. Of course, Unix was blamed. Turned out the print jobs were being sent to an undersized NT print server that served the entire division and it just choked on the end-of-month inundation of print jobs.
 
The following is from iostat 1. If I read this correctly, the drive cltld3 is busiest. Now, how do I know what file systems I have for this drive. I am running HP-UX. The command bdf or /etc/fstab don't seem to relate to cltld3

device bps sps msps
c2t6d0 73 20.3 1.0
c6t2d0 64 13.7 1.0
c6t1d0 2 0.7 1.0
c1t1d3 877 111.1 1.0
c0t0d3 0 0.0 1.0
c1t1d4 1 0.7 1.0
c0t0d4 0 0.0 1.0
c6t13d0 0 0.0 1.0
c1t1d0 0 0.0 1.0
c1t1d1 0 0.0 1.0
c0t0d1 0 0.0 1.0
c0t0d2 5 0.7 1.0
c1t1d2 0 0.0 1.0
c1t1d5 0 0.0 1.0
c1t1d6 5 0.7 1.0
c0t0d6 0 0.0 1.0

bdf:
Filesystem kbytes used avail %used Mounted on
/dev/vg00/lvol3 143360 49988 87555 36% /
/dev/vg00/lvol1 83733 28795 46564 38% /stand
/dev/vg00/lvol8 819200 669567 141229 83% /var
/dev/vg00/lvol7 819200 542641 259278 68% /usr
/dev/vg01/tuxedo 204800 25478 168143 13% /tuxedo
/dev/vg00/lvol6 774144 403292 348615 54% /tmp

/etc/fstab
/dev/vg00/lvol3 / vxfs delaylog 0 1
/dev/vg00/lvol1 /stand hfs defaults 0 1
/dev/vg00/lvol4 /home vxfs delaylog 0 2
/dev/vg00/lvol5 /opt vxfs delaylog 0 2
/dev/vg00/lvol6 /tmp vxfs delaylog 0 2
/dev/vg00/lvol7 /usr vxfs delaylog 0 2
/dev/vg00/lvol8 /var vxfs delaylog 0 2

Thanks,
 
you can do a pvdisplay -v /dev/dsk/c1t1d3 | more to get a very long list of the logical volumes on the disk. You can also use sam to get a list of the LVs on the disk. (Disks and Fielsystems/Disk Devices).

I'm wondering why you don't have your /tuxedo filesystem in your /etc/fstab? Is this an MCServiceGuarded app?

 
Sorry for the confusion. I didn't display the whole thing from fstab. Thanks so very much.

/dev/vg00/lvol3 / vxfs delaylog 0 1
/dev/vg00/lvol1 /stand hfs defaults 0 1
/dev/vg00/lvol4 /home vxfs delaylog 0 2
/dev/vg00/lvol5 /opt vxfs delaylog 0 2
/dev/vg00/lvol6 /tmp vxfs delaylog 0 2
/dev/vg00/lvol7 /usr vxfs delaylog 0 2
/dev/vg00/lvol8 /var vxfs delaylog 0 2
/dev/vg01/nsr /nsr vxfs rw,suid,nolargefiles,delaylog,datainlog 0 2
/dev/vg30/bkdump /bkdump hfs defaults 0 2
/dev/vg31/psinterf /psinterf hfs defaults 0 2
/dev/vg31/oracle /oracle hfs defaults 0 2
/dev/vg01/swap01 ... swap pri=0 0 0
/dev/vg01/swap02 ... swap pri=0 0 0
/dev/vg01/swap03 ... swap pri=0 0 0
/dev/vg01/tuxedo /tuxedo vxfs rw,suid,nolargefiles,delaylog,datainlog 0 2
/dev/vg01/APPS /APPS vxfs rw,suid,nolargefiles,delaylog,datainlog 0 2
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top